Artificial intelligence ai service processing method and device

Abstract

This application relates to an artificial intelligence (AI) service processing method and a device in the field of communication technologies. The artificial intelligence AI service processing method of embodiments of this application includes: receiving, by a network-side device, an AI service request from a target device, where the AI service request includes an AI service identifier of a requested first service and AI assistance information, the AI assistance information includes AI service trigger information and/or assistance information related to an AI model; and feeding back, by the network-side device, a service result of the first service to the target device based on the AI service identifier of the first service and the AI assistance information.

What is claimed is: 
     
         1 . An artificial intelligence (AI) service processing method, comprising:
 receiving, by a network-side device, an AI service request from a target device, wherein the AI service request comprises an AI service identifier of a requested first service and AI assistance information, the AI assistance information comprises AI service trigger information and/or assistance information related to an AI model; and   feeding back, by the network-side device, a service result of the first service to the target device based on the AI service identifier of the first service and the AI assistance information.   
     
     
         2 . The method according to  claim 1 , wherein the assistance information related to the AI model comprises at least one of the following:
 first assistance information for processing input data of the AI model;   input data of the AI model; and   supplementary information of output data of the AI model.   
     
     
         3 . The method according to  claim 1 , wherein the AI service trigger information comprises at least one of the following:
 input data of an AI model of a second service; and   first indication information for indicating validity or invalidity of the input data of the AI model of the second service; wherein   the second service provides input data for the first service.   
     
     
         4 . The method according to  claim 3 , wherein the first indication information comprises at least one of the following: a validity range, an invalidity range, and a valid distance threshold. 
     
     
         5 . The method according to  claim 2 , wherein
 the first assistance information comprises at least one of the following:   second indication information for indicating a validity range or an invalidity range of the input data of the AI model; and   second assistance information for performing coordinate system transformation on the input data of the AI model.   
     
     
         6 . The method according to  claim 5 , wherein the second assistance information comprises at least one of the following: a location of a target object and a timestamp; and the second assistance information is used to assist the network-side device in tracking the target object, wherein the target object is the target device. 
     
     
         7 . The method according to  claim 1 , wherein the feeding back, by the network-side device, a service result of the first service to the target device based on the AI service identifier of the first service and the AI assistance information comprises:
 determining, by the network-side device, an AI service type based on the AI service identifier; and   feeding back, by the network-side device, the service result of the first service to the target device based on the AI service type and the AI assistance information.   
     
     
         8 . The method according to  claim 7 , wherein the AI service type comprises at least one of the following:
 environmental sensing, environmental reconstruction, environmental prediction, vehicle prediction, path planning, and trajectory planning.   
     
     
         9 . The method according to  claim 1 , wherein
 in a case that the first service is environmental sensing, the service result comprises at least one of the following:   traffic light information, obstacle information, or lane line information; wherein   the obstacle information comprises at least one of the following:   an identifier, an obstacle type, an obstacle boundary, an obstacle size, or speed information; and   in a case that the first service is path planning, the service result comprises a plurality of planning information, wherein each planning information comprises at least one of the following:   a trajectory point identifier, a time interval from a current moment, a distance, a speed, and an acceleration.   
     
     
         10 . The method according to  claim 9 , wherein
 the obstacle boundary comprises at least one of the following:   a three-dimensional cuboid boundary with the target object as a center of a coordinate system; and   a two-dimensional rectangular boundary with the target object as the center of the coordinate system; wherein   the target object is the target device.   
     
     
         11 . The method according to  claim 2 , wherein
 in a case that the first service is trajectory planning, the AI service trigger information further comprises first information; or   in a case that the first service is environmental sensing, the first assistance information further comprises first information; wherein   the first information comprises at least one of the following:   a location of a target object; and   a size of the target object;   wherein the target object is the target device.   
     
     
         12 . The method according to  claim 6 , wherein
 in a case that the first service is trajectory planning, the location of the target object is a center of coordinate system transformation.   
     
     
         13 . The method according to  claim 2 , wherein
 in a case that the first service is trajectory planning, the AI assistance information is a sensing result, and the sensing result is input data of an AI model of the first service.   
     
     
         14 . An artificial intelligence (AI) service processing method, comprising:
 sending, by a target device, an AI service request to a network-side device, wherein the AI service request comprises a service identifier of a first service and AI assistance information, and the AI assistance information comprises AI service trigger information and/or assistance information related to an AI model; and   receiving, by the target device, a service result of the first service sent by the network-side device.   
     
     
         15 . The method according to  claim 14 , wherein the assistance information related to the AI model comprises at least one of the following:
 first assistance information for processing input data of the AI model, wherein the first assistance information comprises at least one of the following: second indication information for indicating a validity range or an invalidity range of the input data of the AI model; and second assistance information for performing coordinate system transformation on the input data of the AI model;   input data of the AI model; and   supplementary information of output data of the AI model,   wherein the second assistance information comprises at least one of the following: a location of a target object and a timestamp; and the second assistance information is used to assist the network-side device in tracking the target object, wherein the target object is the target device.   
     
     
         16 . The method according to  claim 14 , wherein
 the AI service trigger information comprises at least one of the following:   input data of an AI model of a second service; and   first indication information for indicating validity or invalidity of the input data of the AI model of the second service, wherein the first indication information comprises at least one of the following: a validity range, an invalidity range, and a valid distance threshold; wherein   the second service provides input data for the first service.   
     
     
         17 . The method according to  claim 14 , wherein
 in a case that the first service is environmental sensing, the service result comprises at least one of the following:   traffic light information, obstacle information, or lane line information; wherein   the obstacle information comprises at least one of the following:   an identifier, an obstacle type, an obstacle boundary, an obstacle size, or speed information, wherein the obstacle boundary comprises at least one of the following: a three-dimensional cuboid boundary with the target object as a center of a coordinate system; and a two-dimensional rectangular boundary with the target object as the center of the coordinate system, and wherein the target object is the target device; and   in a case that the first service is path planning, the service result comprises a plurality of planning information, wherein each planning information comprises at least one of the following:   a trajectory point identifier, a time interval from a current moment, a distance, a speed, and an acceleration.   
     
     
         18 . The method according to  claim 15 , wherein
 in a case that the first service is trajectory planning, the AI service trigger information further comprises first information; or   in a case that the first service is environmental sensing, the first assistance information further comprises first information; wherein   the first information comprises at least one of the following:   a location of a target object; and   a size of the target object; wherein   the target object is the target device.   
     
     
         19 . A target device, comprising at least one hardware processor and a memory, wherein the memory stores a program or instructions capable of running on the at least one hardware processor, and when the program or instructions are executed by the at least one hardware processor, the AI service processing method according to  claim 1  are implemented. 
     
     
         20 . A network-side device, comprising at least one hardware processor and a memory, wherein the memory stores a program or instructions capable of running on the at least one hardware processor, and when the program or instructions are executed by the at least one hardware processor, the AI service processing method according to  claim 14  are implemented.